In large public places such as shopping malls, airports, and train stations, various criminal incidents often occur, including cases such as the elderly or children getting lost, robbery, theft, and other crimes. The harm and pain caused to the victims and their families by these incidents are unimaginable. Therefore, security guards are increased to patrol these public places or surveillance cameras are installed to monitor them.
However, simply using existing surveillance cameras, video management systems, or video surveillance analytics software may achieve partial prevention or monitoring objectives, but there are still many drawbacks. For example, the large number of surveillance cameras in public places, such as shopping malls, airports, or stations, means that security personnel have to retrieve every surveillance video and carefully review the footage from each camera in order to identify suspicious individuals or missing persons manually. This process incurs a significant amount of labor cost and time, and cannot provide real-time and effective crime alerts or missing person assistance.
Therefore, a new generation of AI-powered people tracking technology is integrated into the existing video surveillance analytics system. It can automatically recognize human body shapes and clothing, and track specific individuals based on characteristics such as age, gender, clothing, and accessories. This enhances the effectiveness of the existing video surveillance analytics system, making the entire search and analysis process more precise and efficient.
Video Surveillance Analytics system is a system that uses computer vision technology to capture and analyze images from surveillance cameras to detect, recognize, and analyze people, events, and behaviors. Generally, these systems are also able to store vidoe for long periods of time for subsequent queries and analysis. Video Surveillance Analytics systems have been adopted in various industries, including law enforcement, transportation, retail, healthcare facilities, hotels, and manufacturing, and they provide valuable analysis information to enhance security by detecting threats, preventing crime, and improving situational awareness.
When AI technology is incorporated into video surveillance analytics systems, it can automatically analyze surveillance video without the need for human monitoring and actively alert guards or security personnel to abnormal situations through alarm notifications. Incorporating AI technology can bring several benefits to video surveillance analytics systems, including: higher accuracy, faster response time, savings human resources and costs, ability to handle more complex environmental conditions, and increased sense of security in people's lives.
People Tracking technology is a technique for tracking and analyzing individuals in surveillance footage, typically using AI video analysis algorithms to identify human body contours and features. By tracking and analyzing the movement patterns of contours and features, it can accurately track information such as the selected individual's location, movement speed, and direction. Its main purpose is to identify and analyze the movement patterns of individuals, understand their activity range, movement routes, and behavioral patterns, and further conduct crime prevention, tracking, or searching for missing persons.
People tracking technology can be widely used in fields such as public safety and smart cities. For example, in places with high people traffic such as office buildings, large shopping malls, public spaces, airports, subways, or stations, people tracking technology can be used to enhance public safety and assist in locating missing persons.
Adding people tracking technology to the video surveillance analytics system can further improve the overall system performance. Specifically, there are several applications:
Person re-identification use full-body images to track and find the same person across different cameras. It involves matching individuals in different cameras based on appearance or other features, such as clothing, height, or gait.
In a typical person re-identification system, a camera initially detects and tracks a person, and the system then uses this information to match the same person in other cameras. Due to differences between different camera, as well as the rigid and flexible characteristics of people, appearance can be easily affected by clothing, scale, occlusion, posture, and viewpoint, making people re-identification a challenging and popular topics in the field of computer vision.
To overcome various challenges, person re-identification systems use advanced AI algorithms and technologies such as feature extraction, metric learning, and deep neural networks, which enable the system to extract unique features from a person's appearance and compare them across different cameras. Person re-identification has numerous applications in video surveillance, such as tracking suspects in large public spaces or identifying lost children in crowded areas.
When security personnel do not have a pre-existing image of a person but have knowledge of their attribute characteristics (such as gender, age, clothing, or accessories), person attribute recognition technology can be used to track the person in surveillance footage. Person attribute recognition technology uses deep machine learning to extract features from image and then identifies specific attributes related to individuals (such as gender, age, clothing, or accessories), thereby finding persons in surveillance footage that match those conditions and helping to identify suspects or lost children.
In people tracking technology, facial recognition refers to the technique of identifying people by recognizing their faces captured by cameras.
By integrating FaceMe® facial recognition engine, facial photos can be inputted for more accurate tracking of target individuals. Extraction and matching of facial features of people enable identification and tracking of the same person across different cameras. In practical applications, facial recognition technology can be used to enhance people re-identification, security and other aspects. For example, in surveillance systems, when suspicious behavior occurs, facial recognition technology can be used for more accurate investigation and tracking of potential suspects.
A complete people tracking solution should integrate the three AI technologies mentioned above: person re-identification, person attribute recognition, and facial recognition. Through the integration of these three technologies, people can be identified and tracked more accurately, and further identity information can be obtained. Person attribute recognition technology can identify the appearance characteristics of people, such as gender, age, height, clothing, and accessory information, to initially locate the image of the people to be searched, which can be used by person re-identification technology to track the people under different cameras. Finally, if facial recognition technology is added, the overall people tracking can become more detailed and accurate.
In addition to possessing the three AI technologies mentioned above, a complete people tracking solution should also have the following two features:
CyberLink People Tracker is a new generation smart people tracking solution that perfectly integrates the three major AI technologies mentioned above. It can analyze surveillance camera video in real-time, as well as quickly search for the entry and exit tracks of the target person in surveillance video recordings, greatly enhancing the functionality of existing surveillance and security systems. It also has high accuracy and execution efficiency and is deeply integrated with major mainstream VMS, greatly reducing the time required for tracking.
CyberLink People Tracker is the only solution that perfectly integrates three major AI technologies: facial recognition, person re-identification (Person Re-ID), and person attribute recognition (PAR). This results in highly effective search and tracking capabilities, allowing for tracking of individuals in any pose or angle (walking/running/front/back).
In addition, CyberLink's FaceMe facial recognition technology is a world-leading brand, utilizing deep learning neural network algorithms to achieve up to 99.83% accuracy in recognition (TAR, True Acceptance Rate). Its outstanding recognition capabilities have placed it among the most accurate and fastest recognition engines in the NIST FRVT facial recognition tests, making it the top-ranking facial recognition technology among all global vendors.
CyberLink People Tracker has ultra-high edge computing performance, optimized for AI edge computing accelerators. With a low-end RTX A2000 GPU, it can analyze 21 surveillance cameras simultaneoulsy.
To achieve automatic analysis and recognition of a large number of surveillance video, CyberLink People Tracker is integrated with mainstream video management systems (VMS) to enable users to obtain real-time information about the monitored area. Through this integration, users can easily upgrade their existing surveillance and security systems and enjoy more convenient functions brought by AI, thereby enhancing user and environmental safety.
Smart people tracking solutions can play an important role in ensuring public safety in airports, train stations, shopping malls, tourist attractions, and other public places. For example, in airports and stations, the people tracking solution can help security personnel track passengers suspected of carrying dangerous items, such as large packages, suspicious liquid containers, etc. When a suspicious person is tracked, security personnel can immediately conduct a more detailed security check on that person and take arrest or precautionary measures if necessary.
In shopping malls and tourist attractions, the smart people tracking solution can also play the same role, and even track people with suspicious behavior, such as lingering in dark corners for too long, deliberately following others' movement routes, etc. Through this solution, it can help security personnel quickly identify suspicious individuals and pay attention to them.
In addition, the smart people tracking solution can also be combined with other security systems, such as access control and alarm systems, to achieve a more comprehensive and guarantee public safety intelligently.
Smart people tracking solutions can be applied to various types of security systems in office and residential buildings. In the event of theft or crime in the office or residential environment, the surveillance system can track the suspicious person's movements through AI technologies such as facial recognition, person re-identification, and person attribute recognition. The system can effectively track the suspect's movements inside and outside the building. Once the suspect's movements are identified, the system can immediately notify security personnel or send an alert, further enhancing security effectiveness.
In addition, smart people tracking solutions can also be integrated with smart access control systems to effectively manage the entry and exit of personnel in offices and residential areas through facial recognition and other technologies, further improve security.
When children or individuals go missing in indoor entertainment venues or shopping malls, smart people tracking solutions can quickly extract the missing person's features from surveillance video. By using three technologies, including facial recognition, person re-identification, and person attribute recognition, the system can identify the missing person's age, gender, clothing color, backpack accessories, and other distinguishing features provided by their family members, even in indoor or complex environments. The system can also compare images from different cameras to quickly search and locate the missing person's position, thereby improving search accuracy and efficiency.
Smart people tracking solutions can also be applied in healthcare institutions, nursing homes, and other care facilities to track patients who require special attention, such as those with dementia or mental illness. With this solution, caregivers can instantly detect a patient's situation, such as wandering or falling, and provide timely assistance to reduce the risk of accidents and ensure patient safety.
AI-based people tracking solutions have the potential to revolutionize the existing security and surveillance industry. For example, CyberLink People Tracker is a smart people tracking solution, specific individuals can be recognized and tracked based on age, gender, clothing, and even accessories, even when facial features are not clearly visible. When facial features are available, the solution can track individuals with even greater accuracy. As video analytics algorithms and deep machine learning technologies continue to develop, people tracking technology will become more precise and efficient in the future.
In some environments where facial recognition technology cannot be implemented due to privacy concerns, CyberLink People Tracker can track individuals based on physical features, clothing, or body shape without the need for facial recognition. Going forward, if people tracking solutions are further integrated with other technologies, such as facial recognition, object recognition, and various access control and security systems, they will provide users with a more comprehensive and intelligent security system.